Evaluate statistical properties of McRAPTOR algorithm

conveyal / r5

Developed to power Conveyal's web-based interface for scenario planning and land-use/transport accessibility analysis, R5 is our routing engine for multimodal (transit/bike/walk/car) networks with a particular focus on public transit

https://conveyal.com/learn

MIT License

289 stars 74 forks source link

Evaluate statistical properties of McRAPTOR algorithm #118

Open mattwigway opened 8 years ago

mattwigway commented 8 years ago

We're now using the McRAPTOR algorithm for analysis of Pareto-optimal transit fares and travel time searches. It has a number of things in it that happen randomly (the main one being a constrained random walk to sample over minutes; there will also be randomness happening as a result of doing a Monte Carlo search over the schedules). The results will not be deterministic but we'd like them to have a small confidence interval. We can bootstrap to evaluate this.

mattwigway commented 8 years ago

Well as a first test we appear to be getting convergence, by computing the same result twice using the comparison mode of analyst. This doesn't say much about the frequency of extreme results.

Fare cutoff 1400 COP.

Fare cutoff 2500 COP.